Business Economics
Vol. 49, No. 2
© National Association for Business Economics
U.S. Economic Prospects: Secular Stagnation, Hysteresis, and the
Zero Lower Bound
LAWRENCE H. SUMMERS*
The nature of macroeconomics has changed dramatically in the last seven years. Now, instead of being
concerned with minor adjustments to stabilize about a
given trend, concern is focused on avoiding secular
stagnation. Much of this concern arises from the longrun effects of short-run developments and the inability
of monetary policy to accomplish much more when
interest rates have already reached their lower bound.
This address analyzes contemporary macroeconomic
problems and proposes solutions to put the U.S. economy back on a path toward healthy growth.
Business Economics (2014) 49, 65–73.
doi:10.1057/be.2014.13
Keywords: U.S. economy, stagnation, interest rates,
monetary policy, fiscal policy, economic growth
I
would like to thank Michael Peterson1 very much
for his generous words of introduction and for his
thoughtful observations about the long-run economic
challenges that our country faces. You do not, however,
get to the long run except through the short run, and what
happens in the short run has a profound impact on the
long run. To reverse Keynes a bit, if you die in the short
run, there is no long run. So my preoccupation this
morning will be with a set of temporary but, I believe,
ultimately long-term concerns.
Before I turn to those concerns, however, let me
just say how grateful I am to be back with the National
1
Michael A. Peterson is the Chairman and Chief Operating
Officer of the Peter G. Peterson Foundation.
Association for Business Economics. It seems to me
that the members of this organization make an enormous, ongoing contribution to evaluating, understanding, and responding to the flow of economic events.
I have been coming to these meetings on and off now
for more than 30 years, and I have always been struck
by the sophistication and relevance of the analyses that
are provided herein.
Indeed, I think it is fair to say that some of the
themes that are today central to discussions of academic macroeconomists, but that had receded from the
debate for many years, were always kept alive at the
National Association for Business Economics. I think,
for example, of the importance of the financial sector and
the flow of credit. I also think of the issues surrounding
confidence and uncertainty. These topics have long been
staples of the discussions at NABE meetings.
Macroeconomics, just six or seven years ago, was a
very different subject than it is today. Leaving aside the
set of concerns associated with long-run growth, I think
it is fair to say that six years ago, macroeconomics was
primarily about the use of monetary policy to reduce
the already small amplitude of fluctuations about a
given trend, while maintaining price stability. That was
the preoccupation. It was supported by historical
analysis emphasizing that we were in a great moderation, by policy and theoretical analysis suggesting the
importance of feedback rules, and by a vast empirical
program directed at optimizing those feedback rules.
Today, we wish for the problem of minimizing
fluctuations around a satisfactory trend. Indeed, I think
it is fair to say that today, the amplitude of fluctuations
appears large, not small. As I shall discuss, there is
room for doubt about whether the cycle actually cycles.
Keynote Address at the NABE Policy Conference, February 24, 2014.
*Lawrence H. Summers is the Charles W. Eliot University Professor, the Weil Director of the Mossavar-Rahmani Center for Business
& Government at Harvard’s Kennedy School, and President Emeritus of Harvard University. During the past two decades, he has served in a
series of senior government positions in Washington, D.C., including the 71st Secretary of the Treasury for President Clinton, Director of the
National Economic Council for President Obama and Vice President of Development Economics and Chief Economist of the World Bank.
In 1983, he became one of the youngest individuals in recent history to be named as a tenured member of the Harvard University faculty.
In 1987, Mr. Summers became the first social scientist ever to receive the annual Alan T. Waterman Award of the National Science
Foundation (NSF), and in 1993 he was awarded the John Bates Clark Medal, given every two years to the outstanding American economist
under the age of 40. He received a bachelor of science degree from the Massachusetts Institute of Technology in 1975 and was awarded a
Ph.D. from Harvard in 1982.
Lawrence H. Summers
Today, it is increasingly clear that the trend in growth
can be adversely affected over the longer term by what
happens in the business cycle. And today, there are real
questions about the efficacy of monetary policy, given
the zero lower bound on interest rates.
In my remarks today, I want to take up these
issues—secular stagnation, the idea that the economy
re-equilibrates; hysteresis, the shadow cast forward on
economic activity by adverse cyclical developments;
and the significance of the zero lower bound for the
relative efficacy of monetary and fiscal policies.
I shall argue three propositions. First, as the
United States and other industrial economies are
currently configured, simultaneous achievement of
adequate growth, capacity utilization, and financial
stability appears increasingly difficult. Second, this
is likely to be related to a substantial decline in the
equilibrium or natural real rate of interest. Third,
addressing these challenges requires different policy
approaches than are represented by the current conventional wisdom.
1. The Difficulty of Achieving Multiple Objectives
Let me turn, then, to the first of these propositions.
It has now been nearly five years since the trough of the
recession in the early summer of 2009. It is no small
achievement of policy that the economy has grown
consistently since then and that employment has
increased on a sustained basis. Yet, it must be acknowledged that essentially all of the convergence between
the economy’s level of output and its potential has been
achieved not through the economy’s growth, but
through downward revisions in its potential.
In round numbers, Figure 1 shows that the economy is now 10 percent below what in 2007 we thought
its potential would be in 2014. Of that 10 percent gap,
5 percent has already been accommodated into a
reduction in the estimate of its potential, and 5 percent
remains as an estimate of its GDP gap. In other words,
through this recovery, we have made no progress in
restoring GDP to its potential.
Information on employment is similarly sobering.
Figure 2 depicts the employment/population ratio in
aggregate. Using this relatively crude measure, one
observes almost no progress. It has been pointed out
repeatedly and correctly that this chart is somewhat
misleading because it neglects the impact of a range of
demographic changes on the employment ratio that
would have been expected to carry on even in the
absence of a cyclical downturn.
But that is not the largest part of the story. Even if
one looks at 25-to-54 year-old men, a group where
66
Figure 1. Downward Revision in Potential GDP, U.S.A.
Source: CBO.
there is perhaps the least ambiguity because there is the
greatest societal expectation of work, Figure 3 shows
that the employment/population ratio declined sharply
during the downturn, and only a small portion of that
decrease has been recovered since that time.
The recovery has not represented a return to
potential; and, according to the best estimates we have,
the downturn has cast a substantial shadow on the
economy’s future potential. Making the best calculations one can from the CBO’s estimates of potential
(and I believe quite similar results would come from
other estimates of potential), one can see from Figure 4
that this is not about technological change. Slower total
factor productivity than we would have expected in
2007 accounts for the smallest part of the downward
trend in potential. The largest part is associated with
reduced capital investment, followed closely by
reduced labor input. Let me emphasize that this is not
a calculation about why we have less output today. It is
a calculation about why it is estimated that the potential
of the economy has declined by 5 percent as a
consequence of the downturn that we have suffered.
The record of growth for the last five years is
disturbing, but I think that is not the whole of what
should concern us. It is true that prior to the downturn
in 2007, through the period from, say, 2002 until 2007,
the economy grew at a satisfactory rate. Note that,
there is no clear evidence of overheating. Inflation
did not accelerate in any substantial way. But the economy did grow at a satisfactory rate, and did certainly
U.S. ECONOMIC PROSPECTS
Figure 2. Employment/Population Ratio, Aggregate
Source: U.S. Department of Labor: Bureau of Labor Statistics.
Figure 3. Employment/Population Ratio, Men 25–54
Source: Organization for Economic Co-operation and Development.
achieve satisfactory levels of capacity utilization and
employment.
Did it do so in a sustainable way? I would suggest
not. It is now clear that the increase in house prices
shown in Figure 5 (that can retrospectively be convincingly labeled a bubble) was associated with an
unsustainable upward movement in the share of GDP
devoted to residential investment, as shown in Figure 6.
And this made possible a substantial increase in the
debt-to-income ratio for households, which has been
reversed only to a limited extent, as shown in Figure 7.
It is fair to say that critiques of macroeconomic
policy during this period, almost without exception,
suggest that prudential policy was insufficiently
prudent, that fiscal policy was excessively expansive,
and that monetary policy was excessively loose. One
is left to wonder how satisfactory would the recovery
have been in terms of growth and in terms of
achievement of the economy’s potential with a different policy environment, in the absence of a housing bubble, and with the maintenance of strong credit
standards.
67
Lawrence H. Summers
Figure 4. Why did Potential GDP Fall?
• Potential GDP in 2014
– 2013 estimate vs 2007 estimate: 10% decline
• Why did the estimate decline?
Contribution to
Decline in Estimate
Component of Pot. GDP
Potential TFP
~10% (11%)
Capital
~50% (48%)
Potential Hours Worked
~40% (41%)
Source: CBO data. Author calculations.
Figure 5. Home Prices
Real Home Price Index
200
150
100
50
1900
1950
Date
Source: Robert Shiller’s website.
2000
As a reminder, prior to this period, the economy
suffered the relatively small, but somewhat prolonged,
downturn of 2001. Before that, there was very strong
economic performance that in retrospect we now know
was associated with the substantial stock market bubble
of the late 1990s. The question arises, then, in the last
15 years: can we identify any sustained stretch during
which the economy grew satisfactorily with conditions
that were financially sustainable? Perhaps one can find
some such period, but it is very much the minority,
rather than the majority, of the historical experience.
What about the rest of the industrialized world?
I remember well when the Clinton administration came
into office in 1993. We carried out a careful review of
the situation in the global economy. We consulted with
all the relevant forecasting agencies about the longterm view for global economic growth.
At that time, there was some controversy as to
whether a reasonable estimate of potential growth for
Japan going forward was 3 percent or 4 percent. Since
then, Japanese growth has been barely 1 percent. So, it is
hard to make the case that over the last 20 years, Japan
represents a substantial counterexample to the proposition that industrial countries are having difficulty achieving what we traditionally would have regarded as
satisfactory growth with sustainable financial conditions.
What about Europe? Certainly, for some years
after the introduction of the euro in 1999, Europe’s
economic performance appeared substantially stronger
than many on this side of the Atlantic expected. Growth
appeared satisfactory and impressive. Fears that were
expressed about the potential risks associated with a
Figure 6. Housing Share of GDP
Source: U.S. Department of Commerce: Bureau of Economic Analysis.
68
U.S. ECONOMIC PROSPECTS
Figure 7. Debt/Income Ratio for Households
(Household Debt)/(Disposable Personal Income)
120
Percent
100
80
60
40
1946
1960
1973
1987
2001
2014
Date
Source: Federal Reserve (FRED).
common currency without common governance appeared
to have been overblown.
In retrospect, matters look different. It is now clear
that the strong performance of the euro in the first
decade of this century was unsustainable and reliant on
financial flows to the European periphery that in retrospect appear to have had the character of a bubble. For
the last few years, and in prospect, European economic
growth appears, if anything, less satisfactory than
American economic growth.
In sum, I would suggest to you that the record of
industrial countries over the last 15 years is profoundly
discouraging as to the prospect of maintaining substantial growth with financial stability. Why is this the
case? I would suggest that in understanding this
phenomenon, it is useful at the outset to consider the
possibility that changes in the structure of the economy
have led to a significant shift in the natural balance
between savings and investment, causing a decline in
the equilibrium or normal real rate of interest that is
associated with full employment.
2. The Decline in the Equilibrium Real Rate
of Interest
Let us imagine, as a hypothesis, that this decline in the
equilibrium real rate of interest has taken place. What
would one expect to see? One would expect increasing
difficulty, particularly in the down phase of the cycle,
in achieving full employment and strong growth
because of the constraints associated with the zero
lower bound on interest rates. One would expect that,
as a normal matter, real interest rates would be lower.
With very low real interest rates and with low inflation,
this also means very low nominal interest rates, so
one would expect increasing risk-seeking by investors.
As such, one would expect greater reliance on Ponzi
finance and increased financial instability.
So, I think it is reasonable to suggest that if there
had been a significant decline in equilibrium real
interest rates, one might observe the kinds of disturbing
signs that we have observed. Is it reasonable to suggest
that equilibrium real interest rates have declined?
I would suggest it is a reasonable hypothesis for at
least six reasons, whose impact differs from moment to
moment and probably is not readily amenable to
precise quantification.
First, reductions in demand for debt-financed
investment. In part, this is a reflection of the legacy of
a period of excessive leverage. In part, it is a consequence of greater restriction on financial intermediation
as a result of the experiences of recent years. Yet,
probably to a greater extent, it is a reflection of the
changing character of productive economic activity.
Ponder that the leading technological companies of
this age—I think, for example, of Apple and Google—
find themselves swimming in cash and facing the
challenge of what to do with a very large cash hoard.
Ponder the fact that WhatsApp has a greater market
value than Sony, with next to no capital investment
required to achieve it. Ponder the fact that it used to
require tens of millions of dollars to start a significant
new venture, and significant new ventures today are
seeded with hundreds of thousands of dollars. All of
this means reduced demand for investment, with consequences for equilibrium levels of interest rates.
Second, it is a well known, going back to Alvin
Hansen and way before, that a declining rate of
population growth, as shown in Figure 8, means a
declining natural rate of interest. The U.S. labor force
will grow at a substantially lower rate over the next two
decades than it has over the last two decades, a point
that is reinforced if one uses the quality-adjusted labor
force for education as one’s measure. There is the
possibility, on which I take no stand, that the rate of
technological progress has slowed as well, functioning
in a similar direction.
Third, changes in the distribution of income, both
between labor income and capital income and between
those with more wealth and those with less, have
operated to raise the propensity to save, as have
increases in corporate-retained earnings. These phenomena are shown in Figures 9 and 10. An increase in
inequality and the capital income share operate to
increase the level of savings. Reduced investment
demand and increased propensity to save operate in
the direction of a lower equilibrium real interest rate.
Related to the changes I described before, but I
think separate, is a substantial shift in the relative price
69
Lawrence H. Summers
Figure 8. Population Growth Rate
Annual Growth Rate of US Population
2.0
Percent
1.5
1.0
0.5
0.0
1950
1960
1970
1980
1990
Date
2000
2010
Source: Federal Reserve (FRED).
Figure 9. Corporate Profits
Corporate Profits, as share of GDP
of capital goods. Figure 11 shows the evolution of the
relative price of business equipment. Something similar, but less dramatic, is present in the data on consumer
durables. To take just one example, during a period in
which median wages have been stagnant over the last
30 years, median wages in terms of automobiles have
almost doubled according to BLS data.
Cheaper capital goods mean that investment goods
can be achieved with less borrowing and spending,
reducing the propensity for investment.
Fifth, and I will not dwell on this point, there is a
reasonable argument to be made that what matters
in the economy is after-tax, rather than pre-tax, real
interest rates, and the consequence of disinflation is that
for any given after-tax real interest rate, the pretax real
interest rate now needs to be lower than it was before.
Figure 12 demonstrates this relationship.
Finally, Figure 13 shows that there have been
substantial global moves to accumulate central bank
reserves, disproportionately in safe assets in general,
and in U.S. Treasuries in particular. Each of these
12
Percent
10
Figure 11. Price of capital equipment
8
Ratio of Price Indices: Capital Equipment vs GDP Deflator
1.0
6
Ratio
0.95
4
1950
1960
1970
1980
Date
1990
2000
2010
Source: Federal Reserve (FRED).
0.9
0.85
0.8
1980
1985
1990
1995 2000
Date
2005
2010
(PPI Capital Equip.)/(GDP Price Deflator)
Figure 10. Top 1 Percent
Percent of Aggregate Income
Income Share of Top 1%
20
Figure 12. Inflation, taxes, real interest rates
15
• Consider investor in 40% tax bracket
• Pre-Tax Real Rate = i - π
• Post-Tax Real Rate = (i) (1 - τ) - π
10
0
1940
1960
1980
Year
Source: World Top Incomes Databases.
70
Case 1
(inflation = 3%)
Case 2
(inflation = 1%)
Nominal Rate
5%
1.67%
Pre-Tax Real
Rate
2%
0.67%
Post-Tax Real
Rate
0%
0%
5
1920
2015
2000
2020
U.S. ECONOMIC PROSPECTS
Figure 13. Central Bank Reserves
Notes: Total assets in USD, ratio to nominal GDP in USD. Advanced economies: Australia, Canada, Denmark, the Euro Area,
Japan, New Zealand, Norway, Sweden, Switzerland, the United Kingdom, and the United States. Emerging economies:
Argentina, Brazil, Chile, China, Chinese Taipei, Colombia, the Czech Republic, Hong Kong SAR, Hungary, India, Indonesia, Korea,
Malaysia, Mexico, Peru, the Philippines, Poland, Russia, Saudi Arabia, Singapore, South Africa, Thailand and Turkey. Sources: IMF,
National Data, Haver Analytics & Fulerum Asset Management.
Source: Financial Times.
Figure 14. Natural Rate of Interest
Natural Rate of Interest, from Laubach and Williams (2003)
Percent
6
4
2
the Bank of England, has recently constructed a time
series on the long-term real interest rate on a global
basis, which shows a similar broad pattern of continuing decline.
I would argue first that there is a continuing
challenge of how to achieve growth with financial
stability. Second, this might be what you would expect
if there had been a substantial decline in natural real
rates of interest. And third, addressing these challenges
requires thoughtful consideration about what policy
approaches should be followed.
0
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
Date
Source: Updated estimates from www.frbsf.org/economicresearch/economists/john-williams/.
factors has operated to reduce natural or equilibrium
real interest rates.
What has the consequence been? Laubach and
Williams [2003] from the Federal Reserve established
a methodology for estimating the natural rate of interest. Essentially, they looked at the size of the output
gap, and they looked at where the real interest rate was,
and they calculated the real interest rate that went with
no output gap over time. Their methodology has been
extended to this point, as shown in Figure 14, and it
demonstrates a very substantial and continuing decline
in the real rate of interest.
One looks at a graph of the 10-year TIP and sees
the same picture. Mervyn King, the former governor of
3. Addressing Today’s Macroeconomic
Challenges
So, what is to be done if this view is accepted? As a
matter of logic, there are three possible responses.
Stay patient
The first possible response is patience. These things
happen. Policy has limited impact. Perhaps one is
confusing the long aftermath of an excessive debt
buildup with a new era. So, there are limits to what
can feasibly be done.
I would suggest that this is the strategy that Japan
pursued for many years, and it has been the strategy
that the U.S. fiscal authorities have been pursuing for
the last three or four years. We are seeing very powerfully a kind of inverse Say’s Law. Say’s Law was
the proposition that supply creates its own demand.
Here, we are observing that lack of demand creates its
own lack of supply.
71
Lawrence H. Summers
To restate, the potential of the U.S. economy has
been revised downwards by 5 percent, largely due to
reduced capital and labor inputs. This is not, according
to those who make these estimates, a temporary
decline, but is a sustained, long-term decline.
Reduce the actual real rate of interest
A second response as a matter of logic is, if the natural
real rate of interest has declined, then it is appropriate
to reduce the actual real rate of interest, so as to permit
adequate economic growth. This is one interpretation
of the Federal Reserve’s policy in the last three to four
years. Not in the immediate aftermath of the panic,
when the policy was best thought of as responding to
panic, but in recent years.
This is surely, in my judgment, better than no
response. It does, however, raise a number of questions. Just how much extra economic activity can be
stimulated by further actions once the federal funds rate
is zero? What are the risks when interest rates are at
zero, promised to remain at zero for a substantial
interval, and then further interventions are undertaken
to reduce risk premiums? Is there a possibility of
creating financial bubbles?
At some point, however, growth in the balance
sheet of the Federal Reserve raises profound questions
of sustainability, and there are distributional concerns
associated with policies that have their proximate
impact on increasing the level of asset prices.
There’s also the concern pointed out by Japanese
observers that in a period of zero interest rates or very
low interest rates, it is very easy to roll over loans; and
therefore there is very little pressure to restructure
inefficient or even zombie enterprises. So, the strategy
of taking as a given lower equilibrium real rates and
relying on monetary and financial policies to bring
down rates is, as a broad strategy, preferable to doing
nothing, but comes with significant costs.
Raise demand
The preferable strategy, I would argue, is to raise the
level of demand at any given rate of interest—raising
the level of output consistent with an increased level of
equilibrium rates and mitigating the various risks
associated with low interest rates that I have described.
How might that be done? It seems to me there are a
variety of plausible approaches, and economists will
differ on their relative efficacy. Anything that stimulates demand will operate in a positive direction from
this perspective. Fiscal austerity, from this perspective,
is counterproductive unless it generates so much confidence that it is a net increaser of demand.
72
There is surely scope in today’s United States for
regulatory and tax reforms that would promote private
investment. Although it should be clear from what I am
saying that I do not regard a prompt reduction in the
federal budget deficit as a high order priority for
the nation, I would be the first to agree with Michael
Peterson and his colleagues at the Peter G. Peterson
Foundation that credible long-term commitments
would be a contributor to confidence.
Second, policies that are successful in promoting
exports, whether through trade agreements, relaxation
of export controls, promotion of U.S. exports, or resistance to the mercantilist practices of other nations when
they are pursued, offer the prospect of increasing
demand and are responses to the dilemmas that I have
put forward.
Third, as I’ve emphasized in the past, public
investments have a potentially substantial role to play.
The colloquial way to put the point is to ask if anyone is
proud of Kennedy Airport, and then to ask how it is
possible that a moment when the long-term interest rate
in a currency we print is below 3 percent and the
construction unemployment rate approaches double
digits is not the right moment to increase public
investment in general—and perhaps to repair Kennedy
Airport in particular.
But there is a more analytic case to make, as well.
This will be my final set of observations. With the help
of David Reifschneider, who bears responsibility for
anything good you like in what I am about to say, but
nothing that you do not like, we performed several
simulations of the standard Federal Reserve macroeconometric model—including the version that he,
Wascher, and Wilcox have studied—to address issues
associated with hysteresis coming from the labor
market. To be clear, this is the Federal Reserve model
as it stands, not modified in any way to reflect any
views that I have.
The simulations performed addressed a 1 percent
increase in the budget deficit directed at government
spending maintained for five years, tracking carefully
the adverse effects on the impacts on investment and
labor force withdrawal, which in turn affect the economy’s subsequent potential. The simulations also
recognize that until the economy approaches full
employment, it is reasonable to expect that the zero
interest rate will be maintained, and the standard Fed
reaction function is used after that point.
The results of the simulations are shown in
Figures 15 and 16 and reveal what you might expect
them to show: that while the fiscal stimulus is in place,
there is a substantial response, which is greater when
allowance is made for labor force withdrawal effects
U.S. ECONOMIC PROSPECTS
Figure 15. Simulation Output: Real GDP
Figure 17. Simulation Output: Debt/GDP Ratio
Source: Summers and Reifschneider (2014), ongoing
analysis.
Source: Summers and Reifschneider (2014), ongoing
analysis.
Figure 16. Simulation Output: Expenditure Multiplier
Source: Summers and Reifschneider (2014), ongoing
analysis .
than when no such allowance is made. What is perhaps
more interesting is that you see some long-run impact
of the stimulus on GDP after it has been withdrawn.
That is why the potential multiplier can be quite large.
And my final point concerns the impact of this
fiscal stimulus on the debt-to-GDP ratio, shown in
Figure 17. You will note that with or without taking
into account labor force withdrawal, this standard
macroeconometric model indicates that a temporary
increase in fiscal stimulus reduces, rather than
increases, the long-run debt-to-GDP ratio.
Now, there are plenty of political economy issues
about whether it is possible to achieve a temporary
increase in government spending, and so forth. But I
believe that the demonstration that, with a standard
model, increases in demand actually reduce the longrun debt-to-GDP ratio should contribute to a reassessment of the policy issues facing the United States and
push us toward placing substantial emphasis on
increasing demand as a means of achieving adequate
economic growth. This should serve as a prelude to the
day when we can return to the concerns that I think
almost all of us would prefer to have as dominant: the
achievement of adequate supply potential for the U.S.
economy.
Thank you very much.
REFERENCE
Laubach, Thomas and John C. Williams. 2003. “Measuring the
Natural Rate of Interest.” Review of Economics and
Statistics, 85(4): 1063–1070.
73